Developing Subannual Estimates of Health Insurance Coverage from the American Community Survey: Challenges and Promising Next Steps

Developing Subannual Estimates of Health Insurance Coverage from the American Community Survey: Challenges and Promising Next Steps By Robert Santos Sharon K. Long Dean Resnick Douglas Wissoker Genevieve M. Kenney Kathleen Thiede Call1 March 2013 Prepared for: The Assistant Secretary for Planning and Evaluation United States Department of Health and Human Services Acknowledgements: The funding for this project was provided through HHS/ASPE contract HHSP23320095654WC. The opinions and conclusions expressed in this report are those of the authors and do not necessarily represent the views of the Assistant Secretary for Planning and Evaluation, the U.S. Census Bureau, the Urban Institute or its sponsors or trustees. The authors appreciate the time and ideas of the individuals who were contacted over the course of this research, including Robert Groves, Deborah Griffin, Alfredo Navarro, Ralph Folsom, Stephen Blumberg, Robin Cohen, Andy Peytchev, Bob Fay, Adam Safir, Steve Henderson, Heather Contrino, Karen King, Mark Asiala, Michael Beaghen, Steven Riesz, Jana Asher, and Brett O’Hara. 1 Rob Santos, Sharon K. Long, Dean Resnick, Douglas Wissoker, and Genevieve M. Kenney are at the Urban Institute and Kathleen Thiede Call is at the University of Minnesota. i Table of Contents Introduction .................................................................................................................................... 1 I. Study Methodology ..................................................................................................................... 3 II. Findings on Subannual Estimates from the ACS and Other Federal Surveys ............................ 5 III. Review of Key ACS Design Elements of Relevance to Subannual Estimates ............................ 7 IV. Challenges Associated with Developing Subannual ACS Estimates ...................................... 11 V. Next Steps ................................................................................................................................ 13 References .................................................................................................................................... 16 1 Introduction Following the introduction of a question on health insurance coverage in 2008, the American Community Survey (ACS) has increasingly been used as a source for state-level health insurance estimates (e.g., RWJF, 2012 a, b; Long et al., 2010; Lynch et al., 2010; Boudreaux et al., 2011; Kenney et al., 2011; Mancini et al., 2011; Kenney et al., 2012; Rodean, 2012; CMS 2012). This reflects a number of key advantages of the ACS, including a survey design that supports state representative estimates for all states (and many substate areas2) and the large size of its public use sample relative to the other national surveys that track health insurance coverage.3 As a result, the ACS yields relatively precise state-level estimates of health insurance coverage for the overall population, for key population subgroups and for substate areas in each state. Consistent with these advantages, the Census Bureau currently recommends that those seeking state-level estimates for a number of measures, including health insurance coverage, rely on the ACS instead of the Current Population Survey (CPS) (DeNavas-Walt et al., 2011). With the recent enactment of the Patient Protection and Affordable Care Act (ACA) and its multiyear implementation schedule, monitoring state-level estimates of health insurance coverage is even more important. The ACA is intended to bring about widespread transformation of the nation’s health care system, with many of the changes being implemented by the states, building on the existing variation in state policies, programs and health systems. In 2009 just prior to the passage of the ACA, the uninsurance rate ranged from 4.2 to 23.8 percent across states, with substantial state-level variation also found in rates of both private and public coverage (American FactFinder). The ACA’s significant expansion of public coverage, the creation of health insurance exchanges, and new subsidies for exchange coverage are expected to lead to a substantial increase in insurance coverage at the national level (CBO, 2011), particularly among states adopting the ACA’s Medicaid expansion. The ACS will likely be a critical component of monitoring efforts, including those by the Assistant Secretary for Planning and Evaluation (ASPE), particularly with regard to state-level estimates of health insurance coverage before and after the implementation of key provisions 2 The ACS is based on an address-based sampling frame that includes households from each county in the country, which improves its representativeness for state and substate estimates. 3 For instance, the ACS public use sample for 2009 included